iOpenShell » Technical questions » Raman frequency calculation in Gaussian03 and Gaussview

Raman frequency calculation in Gaussian03 and Gaussview

Moderators: kadir, krylov, piotr.

Page: 1

Author Post
Member
Registered: Sep 2009
Posts: 4
Location: Iasi
Hello,

I use Gaussian03 and GaussView to calculate frequency for some crystals/minerals. I need to use PBC option, and when i press calculation setup with this option: # freq=raman b3lyp/sto-3g with PBC activated; when Gaussian03 start an error appear.

copy-paste from .LOG file

Quote
[...]

Not enough memory for dipole integrals, short by 11420073 words.
Harris functional with IExCor= 402 diagonalized for initial guess.
RepCel: MaxNCR= 383 NClRep= 383 NMtPBC= 3601.
HarFok: IExCor= 402 AccDes= 0.00D+00 IRadAn= 0 IDoV=1
ScaDFX= 1.000000 1.000000 1.000000 1.000000
Not enough memory to run CalDSu, short by 1000000 words.
Error termination via Lnk1e in d:\programe instalate\Gaussian03W\l401.exe at Tue Sep 22 19:38:26 2009.
Job cpu time: 0 days 0 hours 0 minutes 43.0 seconds.
File lengths (MBytes): RWF= 180 Int= 0 D2E= 0 Chk= 1 Scr= 1
Member
Registered: Aug 2008
Posts: 43
Do you have a %mem card at the top of the file? If so, you should increase this by 15+MW. If you don't have that card, your calculations are using the default (around 6MW), and you should add in a %mem card. Out of curiosity, how much memory does your computer have (that will determine how large you can make %mem)?
Member
Registered: Sep 2009
Posts: 4
Location: Iasi
Hi there, this is my new error:

I use:

%chk=C:\Users\Otilia\Desktop\AMS_DATA.chk
------------------------
# freq=raman b3lyp/6-31g

[...]

Symmetry turned off:
Cannot cope with ghost atoms or with translation vectors.
Stoichiometry C6Ca6O18
Framework group C1[X(C6Ca6O18)]
Deg. of freedom 84
Full point group C1 NOp 1
Standard basis: 6-31G (6D, 7F)
Integral buffers will be 262144 words long.
Raffenetti 2 integral format.
Two-electron integral symmetry is turned off.
318 basis functions, 948 primitive gaussians, 318 cartesian basis functions
150 alpha electrons 150 beta electrons
nuclear repulsion energy 4569.3185984843 Hartrees.
NAtoms= 30 NActive= 30 NUniq= 30 SFac= 1.00D+00 NAtFMM= 60 Big=F
FOutLm= 100.00.
Periodicity: 1 1 1
Max integer dimensions: 13 13 4
PBC vector 1 X= 9.4297 Y= 0.0000 Z= 0.0000
PBC vector 2 X= -4.7149 Y= 8.1664 Z= 0.0000
PBC vector 3 X= 0.0000 Y= 0.0000 Z= 32.2416
Recp vector 1 X= 0.1060 Y= 0.0612 Z= 0.0000
Recp vector 2 X= 0.0000 Y= 0.1225 Z= 0.0000
Recp vector 3 X= 0.0000 Y= 0.0000 Z= 0.0310
Generated k point mesh (from -Pi to Pi):
K space mesh: X= 30 Y= 30 Z= 10
A half-cell shift: 1
Using k point mesh (from -Pi to Pi):
K space mesh: X= 30 Y= 30 Z= 10
A half-cell shift: 1
CountK=T Total number of k points: 0
CountK=T Total number of k points: 4500
[b]Out-of-memory error in routine STVDrv-2 (IEnd= 365309167 MxCore= 6291456)
Use %mem=349MW to provide the minimum amount of memory required to complete this step.[/b]
Error termination via Lnk1e in c:\program files\G03W\l302.exe at Sat Oct 03 00:26:50 2009.
Job cpu time: 0 days 0 hours 1 minutes 21.0 seconds.
File lengths (MBytes): RWF= 7 Int= 0 D2E= 0 Chk= 1 Scr= 1


I have 2 GB of RAM (DDR2).
When I use %mem=349MW, Gaussian die (don't work).
The 349 value is equal with 3,5 GB of memory? My PC need a upg to 4 GB for this calc.?
Administrator
Registered: Aug 2009
Posts: 43
Location: Cleveland, OH
Memory upgrade might not help either. There use to be a 256mb memory limit (regardles of amount specified) for Gaussian under Windows environment. However, I am not sure if it still the case. Did you tried to run this job on a linux box? Its only 318 basis functions. BTW, on your (and any other) 32 bit system the maximum amount of memory will be ~3.5gb, even if you install 4gb.
Member
Registered: Sep 2009
Posts: 4
Location: Iasi
Hello there,
I need to run Gaussian under linux platform and this job? I don't tried to run this job on linux platform.
It's works 100%?
Administrator
Registered: Aug 2009
Posts: 43
Location: Cleveland, OH
All serious calculations done under *nix environment. Certainly, you will need better hardware:-)
madrine
Guest
Olexandr Isayev wrote
Memory upgrade might not help either. There use to be a 256mb memory limit (regardles of amount specified) for Gaussian under Windows environment. However, I am not sure if it still the case. Did you tried to run this job on a linux box? Its only 318 basis functions. BTW, on your (and any other) 32 bit system the maximum amount of memory will be ~3.5gb, even if you install 4gb.

It is always a pleasant experience in knowing about the different things said by Olexandr Isayev. I agreed with this thoughts. It is very useful to understand the different things about the calculation of the frequency. That is a very useful experience for anyone in knowing about Gaussian03 and Gauss view.

It will definitely depend upon the views of the people to share their good experience easily. This will actually gives the good thoughts for the people to get it.
Member
Registered: Aug 2013
Posts: 2
I am trying to draw a simple crystal structure of Ca2N using Gaussian09 and GaussView 5.

I have taken a DFT PBEPBE and a basis set of 631-G. And assigning a memory as required by the system (I have actually tried several memory allocation options). But everytime, I am getting an error like:

Not enough memory for dipole integrals, short by 112306552 words.
Dimensions of / B2/ exceeded in AtP2B2.

I have been stuck with this for quite sometime and need some help. I am just starting off with Gaussian and finding my feet.

Thanks in advance.
« Last edit by rayshaumik on Tue Aug 13, 2013 10:56 am. »

Page: 1

iOpenShell » Technical questions » Raman frequency calculation in Gaussian03 and Gaussview